基于YOLOv5算法的无人机巡检电网绝缘子识别研究  

Research on Insulator Identification of UAV Patrol Inspection Network Based on YOLOv5 Algorithm

作　　者：陶思恒 钱懿如 杨易 TAO Siheng;QIAN Yiru;YANG Yi(Guangxi Power Grid Co.,Ltd.,Nanning Guangxi 530015,China;Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen Guangdong 518000,China;Power Transmission Management of Haikou Power Supply Bureau of Hainan Power Grid Co.,Ltd.,Haikou Hainan 570100,China)

机构地区：[1]广西电网有限责任公司,广西南宁530015 [2]深圳供电局有限公司,广东深圳518000 [3]海南电网有限责任公司海口供电局,海南海口570100

出　　处：《信息与电脑》2022年第21期95-97,101,共4页Information & Computer

摘　　要：由于常规电网绝缘子识别方法不能有效提高识别效率,导致难以适应高质量的运维需求,研究基于YOLOv5算法的无人机(Unmanned Aerial Vehicle,UAV)巡检电网绝缘子识别方法。通过电网无人机巡检勘查,结合绝缘子数据集特点对无人机航拍图片进行训练,采取YOLOv5算法准确识别电网绝缘子。实验结果表明,在无人机航拍图像处理并输入网络运算后,基于YOLOv5算法得到的电网绝缘子识别准确率最高,验证了基于YOLOv5算法的无人机巡检电网绝缘子的准确识别能力。Since the conventional grid insulator identification method cannot effectively improve the identification efficiency,which makes it difficult to adapt to the high-quality operation and maintenance needs,the Unmanned Aerial Vehicle(UAV)inspection grid insulator identification method based on YOLOv5 algorithm is studied.Through the grid UAV inspection survey,the UAV aerial pictures are trained by combining the insulator data set characteristics,and the YOLOv5 algorithm is adopted to accurately identify the grid insulators.The experimental results show that after the UAV aerial images are processed and input into the network operation,the grid insulator recognition accuracy obtained based on the YOLOv5 algorithm is the highest,which verifies the accurate recognition capability of the UAV inspection grid insulator based on the YOLOv5 algorithm.

关 键 词：YOLOv5算法 无人机(UAV)巡检 电网绝缘子识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]